Can not use "saveCompactModel" command for k-NN model (fitcknn) in version 2017a

Question

Le Truong An 2019-7-13

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/471542-can-not-use-savecompactmodel-command-for-k-nn-model-fitcknn-in-version-2017a

编辑： Le Truong An 2019-7-13

IrisFlower.mat

Hi everyone,

I couldn't use "saveCompactModel" command for k-NN model (fitcknn) in version 2017a. I think it's available for version 2019a.

Now, I using version 2017a. How to save k-NN model and then check k-NN model with a new dataset?

Could Someone help solve my problem? Pls and Thanks!

Error message!

Undefined function 'toStruct' for input arguments of type 'ClassificationKNN'.

Error in saveCompactModel (line 17)

compactStruct = toStruct(compactObj); %#ok<NASGU>

Error in simpleKNN (line 31)

saveCompactModel(classificationKNN,'mykNN');

This is my code:

clear all;
%% preparing data
load('IrisFlower.mat');
output = grp2idx(Y); 
rand_num = randperm(size(Xnew,1));
% training 70% testing 30% 
X_train = Xnew(rand_num(1:round(0.7*length(rand_num))),:);
Y_train = output(rand_num(1:round(0.7*length(rand_num))),:);
X_test = Xnew(rand_num(round(0.7*length(rand_num))+1:end),:);
y_test = output(rand_num(round(0.7*length(rand_num))+1:end),:);
%% Train a classifier
% This code specifies all the classifier options and trains the classifier.
classificationKNN = fitcknn(...
    X_train, ...
    Y_train, ...
    'Distance', 'Euclidean', ...
    'Exponent', [], ...
    'NumNeighbors', 3, ...
    'DistanceWeight', 'Equal', ...
    'Standardize', true);
%% Compute validation accuracy
% Perform cross-validation
partitionedModel = crossval(classificationKNN, 'KFold', 10);
% Compute validation predictions
[validationPredictions, validationScores] = kfoldPredict(partitionedModel);
% Compute validation accuracy
validationAccuracy = 1 - kfoldLoss(partitionedModel, 'LossFun', 'ClassifError');
%% save model and check model with new dataset
saveCompactModel(classificationKNN,'mykNN');